Introduction

One of the basic handicaps of the fossil record is that it is incomplete and biased. With respect to the record of dinosaurs it is clear from the literature that there is a strong bias toward fossil remains found in the lower latitudes while the higher latitudes, relatively speaking, are underrepresented (Weishampel et al. 2004). Now that a growing number of dinosaur discoveries have come to light in the higher latitudes (de Lapparent 1962; Heintz 1963; Roehler and Stricker 1984; Brouwers et al. 1987; Davies 1987; Parrish et al. 1987; Nelms 1989; Clemens and Nelms 1993; Pasch and May 1997; Gangloff 1995; 1998; Gangloff et al. 2005; Rich et al. 2002; Fiorillo and Gangloff 2000; 2001; Fiorillo and Parrish 2004; Fiorillo 2004; 2006a; Fiorillo et al. 2009; Fiorillo et al. 2010) it is also clear that rather than these remains representing oddities from remote, often cold places, these discoveries provide insights into the dynamics of life in ancient high-latitude terrestrial ecosystems.

The first discoveries of dinosaurs from high latitudes presented challenges to our understanding of dinosaur physiology (de Lapparent 1962; Heintz 1963). To accommodate this apparent paradox some workers have invoked migration models for high-latitude dinosaurs (Heintz 1963; Hotton 1980; Brouwers et al. 1987; Parrish et al. 1987; Currie 1989), with modern caribou (Rangifer tarandus) most commonly used as the modern analog. A more thorough evaluation of the caribou migration model revealed several problems resulting in Fiorillo and Gangloff (2001) rejecting the analogy based on biomechanical grounds and arguing instead that Arctic hadrosaurs lived year-round in the high latitudes. Similarly, southern hemisphere dinosaurs are also now considered year-round residents of the southern high latitudes (Bell and Snively 2008).

If prey did not migrate, their predators need not have either. Examination of the theropod fauna from the Late Cretaceous of northern Alaska showed that one small theropod, Troodon formosus, a taxon that is rare in the more southern latitudes, is overwhelmingly abundant in Alaska (Fiorillo and Gangloff 2000; Fiorillo 2008). Following the work of Russell and Seguin (1982) in which they suggested the taxon was a crepuscular or nocturnal predator, Fiorillo and Gangloff (2000) suggested Troodon formosus was well adapted to thrive in the extremely seasonal light régime present in the Cretaceous high latitudes. Further, this adaptive advantage may have resulted in an increase in overall body size for Alaskan forms of Troodon (Fiorillo 2008).

From this earlier work it now appears that Alaska dinosaurs lived year- round in the high latitudes, and that at least some aspects of the fauna adapted to life in the ancient Arctic. Numerous studies have suggested higher mean annual temperatures in the Cretaceous high latitudes compared to the modern Arctic (Otto-Bliesner and Upchurch 1997; Parrish and Spicer 1988; DeConto et al. 1999; Takashima et al. 2006) but there remains the issue of an extremely seasonal light régime (Spicer 1987, 2003; D'Oro 2006). Further, given that many of the plants of northern Alaska are thought to be deciduous (Spicer 1987, 2003), it is reasonable to assume that there were times of food scarcity. Hence, it is a basic principle in ecology that consumers eat a broader diet when food is scarce, and that consumers specialize when food is abundant (Schoener 1971; MacArthur 1972; Pianka 1978; Emlen 1984; Krebs and Davies 1987; Ricklefs and Miller 2000).

Thus, this paper addresses the issue of dietary preferences of high-latitude dwelling hadrosaurs in relation to the dietary preferences of hadrosaurs found in lower latitudes. The Alaskan form of hadrosaur has traditionally been referred to as Edmontosaurus (Nelms 1989; Gangloff and Fiorillo, 2010). If this identification is correct, this taxon, during the Late Cretaceous, occupied a geographic range extending from modern northern Alaska (Nelms 1989; Gangloff and Fiorillo, 2010) to Baja, Mexico (Weishampel et al. 2004). One modern terrestrial vertebrate has a comparable latitudinal range, the mountain sheep Ovis. Ovis has a geographic range in North America from the Brooks Range in Alaska, extending south through western Canada and the United States to northwestern Mexico (Valdez and Krausmann 1999). Ovis, therefore, provides an analog for how hadrosaurs may have utilized food resources.

Given the extensive area inhabited by sheep, there is a great variety of potential plant species available as food. Nichols and Bunnell (1999), however, identify one feeding generality for the northern Dall and Stone's Sheep; they consume more lichens and mosses than other North American sheep species. Lichens have a high digestibility (amount of food absorbed during consumption) yet they have a low nutritional value (amount of nutrition provided). The harsh environment then forces the sheep to consume food items less than optimal for their needs. In their study of habitat selection for lambing ewes, Rachlow and Bowyer (1998) similarly showed that given the nutritional needs for nursing during lambing, Dall Sheep choose lambing sites that had reduced quantities of moss and lichens, further emphasizing the expanded range of food use during at least non-lambing parts of the annual cycle.

From Ovis one can suggest that northern hadrosaurs may have had a more varied diet than hadrosaurs found in the more southern latitudes. One tool, which has shown tremendous value for inferring dietary patterns in extinct animals, is the analysis of microwear on teeth. The application of this tool to the teeth of hadrosaurs found in Alaska, Alberta, Wyoming, and Texas forms the basis for this discussion.